Self-transformation and structural reconfiguration in coacervate-based protocells

We highlight a new approach for the design and construction of re-configurable soft colloidal scale objects (protocells) based on the pH-induced transition of dipeptide-containing coacervate micro-droplets into discrete aster-like micro-architectures.


Introduction
The design and construction of recongurable so matter is an essential paradigm for understanding the dynamic and adaptive behaviour of living systems. Moreover, recongurable systems that sense and respond structurally and functionally to external stimuli are important for engineering actuated materials and dynamic polymer systems, 1 developing novel strategies in regenerative medicine 2,3 and extending the emerging area of protocell research. 4,5 We recently demonstrated that coacervate micro-droplets prepared by electrostatically induced complexation of counter-charged polyelectrolytes or polyelectrolyte/small molecule aqueous mixtures can be developed as membrane-free, molecularly crowded protocells, 6 and herein we explore the possibility of exploiting these organized micro-ensembles as novel types of so recongurable systems.
Coacervate micro-droplets are produced by liquid-liquid phase separation and exhibit a range of biomimetic properties such as selective molecular uptake, 7,8 micro-compartmentalized nanoparticle 9 or enzyme catalysis, 10 in vitro gene expression, 11,12 and templating of lipid membrane multilayer assembly. 13 Although recent studies have used auxiliary components such as inorganic nanoparticles, 14 inorganic polyanionic clusters, 15 covalent crosslinking, 16,17 and hydrogels 18,19 to produce higher-order coacervate-based micro-architectures, the use of coacervate micro-droplets as an intrinsic, structurally recongurable micro-compartmentalized phase has been rarely exploited. 20 Coacervates based on polymer/small molecule (monomer) complexation show considerable promise as recongurable protocells because the relative weakness of the electrostatic interactions increases the scope to structurally and compositionally manipulate the micro-droplets by environmental triggers such as changes in salt concentration, pH and temperature.
In this paper, we introduce a new strategy for the preparation of coacervate micro-droplets capable of undergoing a pHtriggered process of self-transformation and structural reconguration. For this, we design and prepare a novel polymer-dipeptide coacervate based on the electrostatically mediated complexation of poly(diallyldimethylammonium chloride) (PDDA) and deprotonated N-(uorenyl-9-methoxycarbonyl)-D-Ala-D-Ala (FMOC-AA). In the absence of PDDA, FMOC-AA exists as a monomer in aqueous solution at pH 8, but readily self-assembles into a hydrogel of supramolecular nanobres when the pH is lowered below the pK a of the carboxylic acid group. 21 Signicantly, we exploit this reversible transformation to prepare coacervate micro-droplets that are capable of a pH-induced structural adaptation from a molecularly crowded liquid phase (pH 8.5) to a nanobrous hydrogel network (pH 4.5) (Fig. 1). Our results provide a step towards the assembly of biomimetic micro-droplets exhibiting rudimentary aspects of metamorphosis, and should offer a new approach to the design and construction of so, recongurable micro-ensembles for applications in diverse areas such as sensing, biomedical devices and cell/protocell integration.

Results and discussion
Polymer-dipeptide coacervates were prepared from mixtures of PDDA (average M w ¼ 100 kDa) and FMOC-AA at a range of nal monomer concentrations and PDDA/FMOC-AA monomer molar ratios (Experimental methods). In general, turbid suspensions of micro-droplets were obtained at room temperature and pH 8.5 for mixtures containing 20 mM PDDA monomer units and FMOC-AA concentrations greater than 9 mM (ESI, Fig. S1 †). The mean size of the droplets aer mixing for 1 minute ranged between 200 nm and 5 mm depending on the PDDA monomer/FMOC-AA molar ratio (ESI, Fig. S2, Table S1 †), and was attributed to changes in the rate of coalescence associated with variations in the surface charge of the micro-droplets. Thus, samples prepared at approximately equimolar concentrations or in the presence of excess FMOC-AA were susceptible to coalescence and sedimentation into a bulk phase within 5 to 10 minutes, whilst positively charged PDDA-rich droplets with a typical zeta potential value of +35 mV and mean size of ca. 240 nm remained in suspension even when centrifuged at 16 100 g for 5 minutes. Under close to charge neutral conditions, the concentration of dipeptide within the continuous aqueous phase was ca. 4 mM (ESI, Fig. S3 †), which indicated that 80% of the FMOC-AA was complexed within the coacervate medium to give a local concentration of around 450 mM.
Transformation of the PDDA/FMOC-AA coacervate into a dipeptide hydrogel was achieved by addition of small aliquots of glucono-d-lactone (GDL, nal concentration ca. 20 mM) and leaving the unstirred mixture to age at room temperature (ESI, † Experimental methods). Typically, the polymer-dipeptide micro-droplets transformed into a self-supporting PDDA-containing matrix of bundled supramolecular nanolaments within 24 h ( Fig. 2a and b). The self-structuring process was associated with a slow decrease in pH to a value of 4.5 over a period of 16 h (ESI, Fig. S4 †). Plots of the time-dependent decrease in pH were consistent with previously reported proles for FMOC-dipeptide self-assembly, 22,23 and showed a short intermediate period in which the pH slightly increased or remained constant due to buffering of FMOC-AA via differences in the pK a values of the monomeric and self-assembled forms of the dipeptide. 24 This region of pH invariance, which has been correlated with the onset of FMOC-peptide self-assembly, 23 was delayed in the coacervate medium (t lag ¼ 9.6 AE 2.7 min) compared with self-assembly of FMOC-AA in bulk solution under the same conditions (t lag ¼ 3.9 AE 1.8 min). It seems feasible that electrostatic interactions with the cationic PDDA chains, or the high level of molecular crowding, or both, were responsible for reducing the rate of FMOC-AA deprotonation and hence curtailing the onset of nanolament nucleation. Given the non-covalent nature of the interactions responsible for both liquid-liquid micro-phase separation and nanolament self-assembly, it was possible to transition reversibly between the coacervate and hydrogel phases by appropriate control of the solution pH to regulate the deprotonation/protonation state of FMOC-AA using combinations of GDL/dilute NaOH or gaseous CO 2 /NH 3 (ESI, Fig. S5 †).
We used a range of microscopic and physical methods to assess the properties of the coacervate-derived hydrogels compared with hydrogels prepared by acidication of 20 mM FMOC-AA bulk solutions in the absence of PDDA (ESI, † Experimental methods). The coacervate-derived and FMOC-AA control hydrogels were visually similar, and consisted of entangled dipeptide nanobres comprising laterally bundled protolaments with mean widths of 3.8 AE 0.9 nm and 3.5 AE 0.9 nm, respectively (ESI, Fig. S6 †). In both cases, circular dichroism (CD) spectra of the hydrogels showed bands at 217 nm (n-p*) and 231-303 nm (p-p*) (ESI, Fig. S7 †), indicating a similar superhelical arrangement of alanine and uorenyl residues within the dipeptide nanolaments. 21 Differential scanning calorimetry (DSC) proles showed broad melting transitions at around 72 or 55 C for hydrogels prepared by coacervate-derived transformation or in bulk solution, respectively (Fig. 2c). The increased gel-to-sol transition temperature of the coacervate-derived FMOC-AA hydrogel was attributed to increased stabilization of the dipeptide nanobundles due to favourable interactions with the PDDA chains. On the other hand, the presence of PDDA reduced the mechanical strength of the hydrogel to shear-induced strain compared with analogous materials prepared in bulk solution. In particular, frequency sweeps in the linear viscoelastic region gave elastic moduli (G 0 ) values at 10 rad s À1 of approximately 510 and 4300 Pa, and corresponding loss factors (tan d) of 0.288 and 0.277 for the coacervate-derived and control hydrogels, respectively (Fig. 2d), indicating that the former was less solid-like. 25 However, the The above results demonstrate that by using a structurally adaptive pH-responsive functionalized dipeptide it is possible to prepare coacervates capable of undergoing triggered processes of self-transformation and reconguration. To further elucidate these processes, we used epiuorescence and confocal uorescence microscopy to monitor the time-dependent structural and morphological changes associated with individual PDDA/FMOC-AA micro-droplets whilst undergoing transformation. The droplets (1-30 mm in size) were prepared at a PDDA : FMOC-AA monomer molar ratio of 1 : 1, and stained prior to addition of GDL with the peptide nanobre-binding blue uorescent dye, Hoechst 33258. Time-dependent optical microscopy images showed a progressive roughening in the  texture of the coacervate micro-droplets within a few minutes of GDL addition (Fig. 3a), and corresponding confocal microscopy images showed the emergence of a corona of brous outgrowths that emanated in all directions from the surface of individual PDDA/FMOC-AA coacervate droplets (Fig. 3b). This phenomenon of outward bre growth from coacervate droplets was attributed to the slow hydrolysis rates of GDL and consequential protonation rates of FMOC-AA (ESI, Fig. S4 †). If mineral acids were used to lower the pH, such as HCl, aggregates formed (ESI, Fig. S9 †) comprising ring-like structures of FMOC-AA around coacervate droplets, which was discerned by the spatial localisation of Hoechst 33258 uorescence.
Fluorescence microscopy images indicated that Hoechst 33258 was homogeneously sequestered into the coacervate droplets prior to addition of GDL (Fig. 4a). On addition of GDL, and aer an induction period that depended on the size, number and spacing of the micro-droplets present on the microscope slide, individual coacervate droplets became surrounded by an aster-like corona of densely packed short bres, which exhibited high intensity blue uorescence (Fig. 4b). The stained images were consistent with binding of the dye to a supramolecular nanolamentous assembly of FMOC-AA molecules. Continued growth of the nanobres produced individual droplets with a contracted core enclosed within an extended mesh of highly elongated FMOC-AA bres (Fig. 4c). The dipeptide bres were unbranched, exible, relatively uniform in width, considerably longer than the coacervate droplet core, and self-limiting with regard to their maximum extension. We attributed the latter to hydrogel formation within the core region, which curtailed nanobre growth by depleting the concentration of protonated FMOC-AA molecules released at the surface of the transforming coacervate micro-droplets. Signicantly, we doped the coacervate mixture with 1 mol% of rhodamine B isothiocyanate (RITC)-labelled poly(allylamine hydrochloride) (PAH), and used confocal uorescence microscopy to determine the spatial distribution of the cationic polymer within the transforming droplets. The images showed that the polymer was specically located in the core of the spherulitic structure and not associated strongly with the emanating FMOC-AA bres (Fig. 4d). In certain circumstances, local alignment of the adjacent dipeptide bres around the surface of a single coacervate core produced a central ring-like bundle of laments that was retained in the aged hydrogels ( Fig. 4e and ESI, Fig. S10 †). Analysis of video images recorded on individual bres during the initial stages of outgrowth (ESI, Fig. S11 †) indicated that the rate of nanolament assembly followed sigmoidal kinetics (Fig. 4f). A corresponding histogram of the maximum growth rates (V max ) of individually tracked dipeptide bres showed a bimodal distribution comprising two populations with V max values of 0.4 AE 0.16 and 0.7 AE 0.08 mm s À1 (Fig. 4g), which were attributed to the growth of bundled co-aligned bres and single laments, respectively.
Given the above observations, we were able to recapitulate formation of the polymer-peptide coacervate micro-droplets by pH-induced transformation of the aster-like dipeptide structures. For this, we added equimolar amounts of sodium hydroxide to counter the addition of GDL, de-protonate FMOC-AA, and reinstall electrostatic interactions between the cationic polymer and functionalized amino acid. Re-formation of the coacervate micro-droplets occurred almost instantaneously upon addition of hydroxide, and was dependent on the hydroxide ion diffusion gradient produced on injection of the alkaline solution (Fig. 5a, ESI, Movie S1 †). In general, two main re-assembly pathways were observed involving local retraction of the aster-like structures back into a single coacervate droplet (Fig. 5b), or division into multiple droplets (Fig. 5c, ESI, Movie S2 †). We attributed the uncontrollable ssion mechanism to turbulent ow associated with the sodium hydroxide gradient, which sheared the nanobrous structure and induced re-coacervation at multiple sites to produce several daughter droplets. Interestingly, the pH-induced ssion process offers a possible route to coacervate micro-droplet division, which we will explore in future work.
The inuence of coacervate composition on the self-structuring process was investigated by preparing PDDA/FMOC-AA micro-droplets at lower FMOC-AA concentrations or with a lower molecular weight (8.5 kDa) PDDA polymer. In general, decreasing the monomer molar ratio from 1 : 1 (as described above) to a value of 1 : 0.85 reduced the rate of nucleation and outgrowth of the dipeptide brous shell, which in turn produced individual coacervate droplets surrounded by a dense brush-like corona with a spiral texture ( Fig. 6a and ESI,  Fig. S12a †). Lowering the PDDA/FMOC-AA molar ratio to 1 : 0.5 greatly prolonged the onset of bre outgrowth to approximately 1 h, and produced a relatively thick, homogeneous shell in which the individual dipeptide bres could not be resolved by uorescence microscopy (Fig. 6b and ESI, Fig. S12b †). The slow decomplexation-mediated release of protonated FMOC-AA molecules observed under these conditions suggests that the dipeptide is more strongly bound within the coacervate matrix in the presence of excess PDDA. As a consequence, nucleation and growth of the dipeptide bres specically at the droplet surface are less competitive compared with bre self-assembly arising from released FMOC-AA molecules present in the bulk solution, such that only short, non-distinct bres are produced in the coronal layer.
Similar experiments were undertaken with polymer/dipeptide coacervate droplets prepared at a molar ratio of 1 : 1 but using a PDDA polymer with an average molecular weight of 8.5 kDa in place of the 100 kDa polymer employed in the experiments described above. Use of the shorter chain polycation was expected to destabilize the coacervate matrix by reducing the attractive interactions between the polymer and FMOC-AA components. As a consequence, although nanobre aster-like microstructures were readily produced under these conditions, the emanating dipeptide brous bundles were thinner, less well-dened, and more branched than those observed in the presence of the 100 kDa polymer ( Fig. 6c and ESI, Fig. S12c †). The change in morphology was consistent with the observed faster kinetics of droplet transformation, and indicated that self-assembly of the dipeptide bres specically on the droplet surface was highly competitive when compared with nucleation and growth in free solution.
Given the highly anisotropic nature of the dipeptide outgrowths associated with individually transforming coacervate droplets, we also undertook preliminary studies to determine whether 2D assemblies of PDDA/FMOC-AA (1 : 1) micro-droplets could be exploited for the generation of a self-structured environment of interpenetrating bre networks. For this, we mounted samples of closely spaced coacervate droplets onto PEG-functionalised glass capillary slides and added GDL to initiate droplet transformation. Depending on the separation distance between droplets, the bre outgrowths became intertwined within 30-40 min with neighbouring aster-like microstructures to form an interconnected brous matrix. The degree of entanglement appeared to be controlled by the size of initial droplets, number density within a localized area, and inter-droplet distances ( Fig. 6d and e). 3D stacked images obtained from confocal uorescence microscopy showed a micro-ensemble of interconnecting dipeptide bres emanating from a series of coacervate nodal points and propagating principally along the surface of the supporting glass substrate ( Fig. 6f and ESI Fig. S13 †).

Conclusion
In conclusion, our results indicate that rudimentary aspects of metamorphosis can be integrated into coacervate-based protocells by using a pH-responsive, structure-adaptive dipeptide as a building block of the phase-separated mixture. Transformation of the spherical PDDA/FMOC-AA micro-droplets into discrete aster-like micro-architectures is dependent on a pH diffusion gradient generated by the controlled release of H + ions accompanying the slow hydrolysis of GDL in the continuous phase. Protonation of FMOC-AA results in decomplexation of the coacervate matrix and concomitant self-assembly and outgrowth of a densely packed corona of dipeptide nanobres specically on the surface of individual micro-droplets. In contrast, experiments in which GDL was substituted for a stronger acid (HCl) produced discrete hydrogelled coacervate particles with spherical morphology and non-hairy surface texture. Under these conditions, rapid protonation of the dipeptide molecules results in spontaneous nucleation and assembly of the FMOC-AA nanobres throughout the coacervate micro-droplets rather than outgrowth at the droplet/water interface along an established pH diffusion gradient, indicating that formation of the aster-like structures is under kinetic control. Signicantly, our studies indicate that reconguration of the protocells results in entanglement of the aster-like microstructures and subsequent formation of an interpenetrating reversible brous network that slowly and reversibly transforms into a polymer-containing dipeptide hydrogel. In general, our results suggest that the ability to integrate primitive aspects of structural and morphological transformations in coacervate micro-droplets offers a step towards the design and construction of so, recongurable chemical micro-ensembles, and provides new opportunities for applications in areas such as environmental sensing, biomedical and bio-inspired materials engineering, and storage/release technologies.

Experimental
Preparation and transformation of polymer-dipeptide coacervate micro-droplets Poly(diallyldimethylammonium chloride) (PDDA) with a molecular weight of 100 kDa (ca. 620 monomer units, monomer ¼ 161.7 g mol À1 ) or 8.5 kDa (ca. 50 monomers) was dissolved in water (pH ¼ 8-9) at a monomer concentration of 40 mM. Coacervates were prepared by addition of 100 mL of aqueous N-(uorenyl-9-methoxycarbonyl)-D-Ala-D-Ala (FMOC-AA, 20-40 mM, Bachem) to a 100 mL aqueous solution of PDDA (20-40 mM in monomer) at a nal pH of 8.5. The nal PDDA/FMOC-AA monomer molar ratios were 1 : 1, 1 : 0.85 and 1 : 0.5 (100 kDa PDDA), or 1 : 1 (8.5 kDa PDDA). In each case, the turbid suspension of liquid micro-droplets was centrifuged at 16 100 g for 5 min to produce a bulk continuous coacervate phase and supernatant. The samples were then gently agitated with a plastic pipette to re-suspend the bulk coacervate into micro-droplets that were larger than the primary droplets. Typically, droplet sizes of ca. 3-4 mm were produced, along with a signicant proportion of droplets greater than 10 mm (ESI, Fig. S14 †). The latter were routinely imaged using a range of microscopy techniques.
The concentration of FMOC-AA in the measured volume of the supernatant phase, produced by centrifugation of the PDDA/FMOC-AA coacervates into a bulk phase (pelleted coacervate phase) was determined by UV-vis spectroscopy (FMOC-AA extinction coefficient at 265 nm ¼ 16 553 M À1 cm À1 ). The difference between the total concentration used in the preparation, and concentration determined in the supernatant was employed to determine the FMOC-AA concentration in the measured volume of the bulk coacervate phase.
Transformation of the PDDA/FMOC-AA micro-droplets into dipeptide nanolaments and subsequent extension into a hydrogel network was initiated by addition of 1-2 mL of a 2 M glucono-d-lactone (GDL; nal concentration ca. 20 mM) solution to 200 mL of a coacervate suspension to reduce the pH from 8.5 to around 4. Typically, the initial stages of transformation occurred within a few hours aer adding GDL. Aer leaving these samples to age for 1 day at room temperature, a self-supporting hydrogel was produced. Alternatively, hydrogelation of the polymer-dipeptide coacervates was achieved by addition of CO 2(g) above a freshly prepared suspension of 1 mL PDDA (100 kDa)/FMOC-AA (1 : 1) coacervate solution at a rate of 0.3 L min À1 . The dispersion typically gelled within 2 hours. Reversible re-assembly of the coacervate droplets was achieved by room temperature addition of sodium hydroxide (20 mM, nal pH ca. ¼ 8.5) to the network of peptide nanolaments, or alternatively, by addition of NH 3 vapour from a 35 wt% aqueous ammonia solution placed above the hydrogel for 5 minutes (nal pH of 8.8).

Imaging studies
Transformation of the polymer-dipeptide micro-droplets aer addition of aqueous GDL was studied by epiuorescence or confocal microscopy. Aer mixing, the samples were immediately mounted onto polyethylene glycol (PEG)-functionalised capillary slides and imaged using a Leica DMI3000B inverted microscope with equipped 40Â or 20Â lenses, or a Leica SP8 AOBS confocal laser scanning microscope equipped with a glycerol immersion 63Â lens and an automated shutter control capable of imaging every 5 seconds. Samples were loaded onto PEG-functionalized capillary slides to avoid wetting of the coacervate droplets. Spatial localization of PDDA during the transformation process was determined by doping the cationic polymer with 1 mol% rhodamine B isothiocyanate (RITC)-labelled poly(allylamine hydrochloride) (RITC-PAH; PDDA : RITC-PAH ¼ 50 : 1). Formation of the FMOC-AA nano-laments was tracked by addition of 20 mM of the peptide brebinding dye Hoechst 33258 to the PDDA/FMOC-AA coacervate suspension. Video images were recorded from a time series of confocal microscopy images using the Manual Tracking plugin in Fiji. The scalar distance from the base to the tip of the individually growing nanolaments was determined at 5 s time intervals. The maximum growth rate (V max ) was calculated from individual tracks using a purpose written script in MATLAB®. The script tted a sigmoid function to the tip-to-base displacement over time of an individually tracked lament and calculated the gradient at the inection point to determine V max . Approximately 170 laments were tracked from 20 different coacervate droplets. A kernel density estimate was used to determine the probability density function of the collected growth rates. Log normal functions were tted to the kernel density estimate to calculate the average growth rates of different populations.
The reversibility of transformed nanobrous asters was monitored by epiuorescence microscopy aer addition of sodium hydroxide to PDDA (100 kDa)/FMOC-AA (1 : 1) systems aer 20-30 minutes of GDL addition. Typically 20 mM of sodium hydroxide was added to solutions in capillary slides creating diffusion of sodium hydroxide from the top to the bottom of the sample. As the sodium hydroxide gradient approached aster-like structures, reassembly occurred back to coacervate droplets.

Dipeptide solution preparation and hydrogel formation in bulk solution
Typically, 200 mL of 1 M sodium hydroxide was added in 1-10 mL aliquots to a 4.8 mL aqueous suspension of N-(uorenyl-9methoxycarbonyl)-D-Ala-D-Ala (FMOC-AA) to produce an aqueous solution of the dipeptide (40 mM, nal pH of 8.5). Dissolution of FMOC-AA was facilitated by sonication of the suspension aer addition of each NaOH aliquot using an Ultrawave Q-Series ultrasonication bath. The pH was kept below pH 9 during addition of NaOH to prevent disassociation of the uorenyl group. The FMOC-AA solution was then passed through a 200 mm lter, and used within three days of preparation. FMOC-AA hydrogels were prepared by addition of glucono-d-lactone (GDL, nal concentration, 20 mM) to a 20 mM FMOC-AA aqueous solution at pH 8.5, and le to age for 24 hours before analysis.

Preparation of rhodamine B isothiocyanate-tagged poly(allylamine hydrochloride)
A 10 mg mL À1 solution of poly(allylamine hydrochloride) (PAH) (15 kDa) was prepared in 100 mM of EPPS buffer (4-(2-hydroxyethyl)-1-piperazinepropanesulfonic acid) at pH 9.5. A volume of 2.83 mL of a solution of rhodamine B isothiocyanate (RITC) dissolved in DMSO (1 mg mL À1 ) was added dropwise to 10 mL of the PAH solution. The reaction was incubated overnight in the dark and under constant stirring to produce RITC-labelled PAH. Removal of unreacted RITC and buffer was achieved by dialysis (molecular weight cut-off of 7000 Da) against Milli-Q water over two days with regular changes of the water. The RITC-PAH solution was then lyophilized and stored in the dark prior to use. UV-vis spectroscopy (RITC extinction coefficient, 3 (559 nm) ¼ 62 100 M À1 cm À1 ) was used to determine the RITC : PAH-monomer molar ratio. Typically, the reaction produced a dye-labelled polymer with a RITC : PAH-monomer molar ratio of 1 : 60.

Determination of critical coacervation concentration
The critical coacervation concentration (CCC) associated with formation of the PDDA/FMOC-AA coacervate was determined by monitoring the increase in turbidity of an aqueous solution of PDDA (20 mM, 100 kDa, pH 8.5) on addition of 0.5 mM increments of aqueous FMOC-AA. The turbidity was measured using a Perkin Elmer Lambda 25 UV-vis spectrometer and by monitoring the changes in absorbance (A) at a xed wavelength (l ¼ 500 nm). Turbidity values were calculated from (100 À % transmission (T)) where % T ¼ 100 Â 10 ÀA . The CCC was determined at the point of rapid increase in turbidity associated with formation of the coacervate droplets. The concentration of FMOC-AA in the supernatant phase produced by centrifugation of the PDDA/FMOC-AA coacervates was determined by UV-vis spectroscopy (FMOC-AA extinction coefficient of 3 (265 nm) ¼ 16 553 M À1 cm À1 ). The extinction coefficient was determined by measuring known concentrations of FMOC-AA to determine a linear t.

General methods
Dynamic light scattering (DLS) and zeta potential measurements were performed using a Malvern Zetasizer Nano-ZS equipped with an internal Peltier stage. Hydrodynamic diameters and zeta potentials associated with coacervation were determined 1 min aer mixing of the PDDA and FMOC-AA solutions at pH 8.5. Circular dichroism (CD) spectroscopy measurements on dipeptide hydrogels prepared from aqueous FMOC-AA solutions or PDDA/FMOC-AA coacervates were undertaken at 25 C using a JASCO J-815 spectropolarimeter tted with a Peltier stage. The hydrogels were placed between two quartz plates prior to CD analysis. Transmission Electron Microscopy (TEM) was undertaken on a Jeol TEM 2010 using a LaB 6 lament at 120 keV in bright eld mode. Imaging of FMOC-AA nanolaments was performed by diluting 10 mL of a hydrogel into 100 mL of deionised water followed by mounting 5 mL of the dispersion onto a carbon-coated copper grid and le to dry at room temperature. Negative staining of the peptide nanolaments was achieved by mounting 5 mL of a 1 wt% solution of uranyl acetate onto the TEM grid and then drying with lter paper aer 5 minutes. AFM studies were performed by depositing samples onto freshly cleaved muscovite mica and drying with compressed nitrogen aer set time intervals. AFM imaging was conducted using a Multi-mode VIII microscope utilising Peakforce control (Bruker, USA) in ambient conditions. Cantilevers with spring constant 0.4 N m À1 (Bruker, SCANASYST-AIR-HR) were used with the Bruker high-speed scan head to achieve high resolution imaging of regions up to 100 mm.